TFW you are still using AXPY as a representative workload to make your architectural case in 2019.https://twitter.com/fanf/status/1098171169902088192 …
Do you lean toward SIMD or vector architectures? I’m not an expert here, but there is a nice simplicity to the vector code, if nothing else.
-
-
Thinking of it as either-or is a mistake. Even if you go whole-hog on vector, you want to have "vector-of-SIMD" to support working on vectors of complex, quaternions, interleaved color channels, etc. ARM's SVE is the closest thing to a real proposal that's in public view.
-
Also swept under the rug in Patterson's analysis: most of the ISA explosion in SIMD is integer operations that DO NOT EXIST in scalar form (mul-hi with rounding, etc). You'll still need those in your vector architecture, so you're not actually going to cut down the ISA much.
- 2 more replies
New conversation -
-
-
It think the vector design is interesting but not because it reduces the size of the instruction set, or the number of instructions needed in your hotspot kernels (which are going to be hand-unrolled anyway)
-
Right! The actual wins for a good vector design should come in the not-hot kernels.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.